[tex-live] status report on TL? (LaTeX release)

[Gerben Wierda]

On Saturday, Jul 26, 2003, at 22:24 Europe/Amsterdam, Karl Berry wrote:

> Gerben, I don't mean to imply that I want any sort of restrictions or 
> to
> make any of your work harder.  (Not that policy questions are up to me,
> anyway, of course.)  Just the opposite.
>
> I've lost track of what the real issue is here.  Are there procedural
> changes of some sort that would make things better for you?

It is just that the frequency of TL is a bit less than the frequency of 
important updates. If somebody updates mminstance, it shoul d not take 
1.5 years to range the general public. If pdfTeX gets improved (and 
that happened quite a lot) the general public should profit and not 
have to wait for more than a year. Often the improvements and fixes are 
important enough and I certainly have many users who have no idea about 
compiling your own binaries, libraries and programs.

The way I have been doing this is by following Build/source which has 
been updated with major new releases (when stable) and bugfixes. So it 
has been practice over the last years that Build/source contains a 
"release candidate", something that is stable enough for release. Let's 
say that happens something like once every 3-4 months, but only when 
there is a good reason (important bugfixes/imprvements)

Regularly, when Build/source.development was in a state deemed to be 
stable, Sebastian has updated Build/source. And that enables people 
like me to offer a very up-to-date but also stable update for users.

So, I think a clean setup would be:
	1. Image: what has been released as TL2003, created from Build.source 
at a certain point in time
	2. Build/source: what has "release candidate" status, created from 
Build/source.development at a certain point in time
	3. Build/development: playground

TeX Live has always been used as the basis of a CD release. But these 
days may users also use TeX Live (through me) as the basis of an 
internet-based release. Programs only of course (Master is too big), 
but still rather useful. A well thought-out procedure is good for 
supporting both results out of TeX Live.

G
--
"To be or not to be, that is the question" -- Parmenides